|
|
|
|
|
by lazyjones
5006 days ago
|
|
For much of this frustration I blame the modern practice of letting some committee design a protocol or other standards and then force that down the throats of programmers. This leads to bloated designs and deficient implementations and documentation and last, but not least, far too many revisions and alternatives. Back in the day, protocols (tools, languages, ...) were designed by people who thought very hard about the implementation, the required resources and the programmers whom these were inflicted upon. In contrast, entities like the W3C display complete ignorance for the implementation details (look how they've failed to provide even a half-assed implementation of a browser as proof-of-concept for their "designed by committee" standards). Whatever builds on top of such lackluster work is doomed and will frustrate programmers endlessly ... |
|
For instance, the STEP AP203 standards document is 500+ pages long. (That's the old edition, I don't have a printed copy of the second edition handy.) But those pages don't tell you how to use the hundreds of classes they define to actually create a meaningful file. For that you need the supplemental "recommended practices" document (http://www.steptools.com/support/stdev_docs/express/ap203/re...), which has been marked "preliminary" for the last 14 years and is itself not particularly clear. (On the bright side, at least it's free!)
Of course, that only applies to one of the many different STEP application protocols, and while I did find a reference to an AP214 (which I also need to support) recommended practices document, the link to it was dead.